道路综合特征下高分辨率遥感影像的提取

发布时间：2018-05-20 08:32

本文选题：高分辨率遥感影像 + 纹理特征　；参考：《测绘通报》2017年08期

【摘要】：针对在高分辨率遥感影像中如何提高道路信息提取的准确度和信息量这一问题,通过对影像光谱和纹理特征的分析,将影像特征按照2种光谱特征和3种纹理特征进行分类,进而改善传统的图像分割方法,选择灰度级数和像素对的相对方向、距离和窗口大小作为参数,再通过灰度共生矩阵运算获取影像的纹理信息,通过对这些纹理特征的综合比较分析,最后确定角二阶矩、熵和对比度作为道路纹理特征统计量;再通过对图像像元分析比较,将图像像元标准差和灰度均值作为道路信息提取的光谱特征;在对道路综合特征分析基础上,再通过对遥感图像几何特征分析,最后利用数学形态学的开运算、闭运算、腐蚀、细化等模型算法对遥感图像进行精细化处理,得到道路提取较好的结果。该方法可用于复杂路况的道路信息提取。
[Abstract]:Aiming at the problem of how to improve the accuracy and amount of road information extraction in high-resolution remote sensing images, the image features are classified according to two spectral features and three texture features by analyzing the spectral and texture features of the images. Then the traditional image segmentation method is improved, the relative direction, distance and window size of gray series and pixel pair are selected as parameters, and then the texture information of image is obtained by gray level co-occurrence matrix operation. Through the comprehensive comparison and analysis of these texture features, the angular second-order moments, entropy and contrast are determined as statistics of road texture features. The image pixel standard deviation and gray mean are taken as the spectral features of road information extraction. Based on the analysis of road synthetic features, the geometric features of remote sensing images are analyzed, and finally, the mathematical morphology is used to open, close and corrode. The fine processing of remote sensing image by thinning and other model algorithms is carried out, and the better result of road extraction is obtained. This method can be used to extract road information from complex road conditions.
【作者单位】：辽宁地质工程职业学院;东北大学资源与土木工程学院;
【基金】：国家自然科学基金青年项目(41104104)
【分类号】：P237

【相似文献】